New Peak Finding Algorithm for the BCM1F Detector of the CMS Experiment at CERN

The Compact Muon Solenoid experiment (CMS) is located at the most powerful particle collider in the world, the Large Hadron Collider (LHC). It detects the resulting prod- ucts of the proton beams that have been accelerated and collide in the center of the experiment. For the operation of a particle collider, a prompt measurement result of the so-called luminosity is of high priority, as it quantifies the ability of the collider to produce a certain number of specific particle interactions. The BCM1F detector is one of the instruments that are installed along the beam pipe to provide an online measure- ment of the luminosity. For reliable results, the electrical pulses that are produced by particle hits at the sensors have to be detected and characterised by their relative arrival time. In the currently operating back end system, the pulses are detected using a single threshold discriminator. This approach shows hit counting inefficiencies, especially for closely occurring, consecutive pulses. A similar pulse detection implementation for the new BCM1F back end system shows even higher inefficiency. This thesis explores the peak finding capability of an algorithm that finds pulse peaks by using the derivative of the signal. The different approaches are compared and the advantages of the derivative based algorithm are explained. The first chapter introduces the reader to the Large Hadron Collider and the Com- pact Muon Solenoid experiment, as well as LHC specific terminology and quantities. Chapter 2 sets the purpose of this thesis into a context and explains the motivation of this work. The concept and details of luminosity, as well as the luminosity measurement at the CMS experiment are explained in chapter 3. The architecture and working prin- ciple of the BCM1F detector are explained in detail in chapter 4. The building blocks of BCM1F from the front end to the back end are characterised in chapter 5, to create a software model of the detector signal. In chapter 6 the discriminator based approach of the operating system and the current peak finder implementation are introduced and their inefficiencies are discussed. Chapter 7 then presents the new derivative based peak finding algorithm and its implementation in VHDL. The performance is evaluated and compared to the discriminator based approaches in chapter 8. The conclusion about the implemented derivative based peak finding algorithm is drawn in chapter 9, followed by an outlook and the perspective of the peak finder application and possible improvements in chapter 10.